Cocaine, methylphenidate and secobarbital were compared on a drug maintained progressive-ratio procedure in baboon subjects. Trials, scheduled throughout the day, occurred at a minimum interval of 3 hrs after completion of the preceding trial. A ratio response requirement on the "initiate" lever was required during each trial which terminated in a single intravenous infusion of drug. A drug was introduced on the progressive-ratio procedure with a low ratio requirement in order to obtain a baseline performance of a high stable frequency of trial completion. The ratio requirement was systematically increased every 7 days until the "breaking point" when the rate of completing trials fell below a criterion level. Within-subject comparison revealed that cocaine produced higher breaking points than methylphenidate at the same absolute dose, 0.4 mg/kg. At the range of doses studied, manipulation of doses of methylphenidate (0.1-0.8 mg/kg) and cocaine (0.4-1.6 mg/kg) had little effect on breaking point. In contrast, increasing doses of secobarbital (6.0 and 12.0 mg/kg) produce higher breaking points within the same subjects.